Published under the title "Poverty, Race, and the Contexts of Achievement: Examining Educational Experiences of Children in the U.S. South"

"This article considers issues of educational inequality in the U.S. South from a social work/social justice perspective. After a review of existing literature and discussion of cultural versus structural explanations for race and socioeconomic status gaps in academic achievement, findings are presented from a study examining child-, classroom-, and school-level factors that influence academic achievement among public school children in the South. Although a sizeable minority of southern children attend schools that are segregated along racial and socioeconomic lines, and although these schools are different in various aspects of educational environment, once family structure, parental characteristics, the use of ability grouping, and rural school location were taken into account, no influence of race on achievement remained.

Implications for social work policy and practice are discussed.

Race and socioeconomic status gaps in children's academic achievement are a troubling social justice issue, both because of the serious long-term social and economic consequences, and because despite decades of research and efforts at reform, these gaps have proven quite robust (Braun, Wang, Jenkins, & Weinbaum, 2006). Achievement gaps are a particularly critical social work issue in the American South-an area characterized by high levels of poverty (Rural Poverty Research Institute, 2001), a large black population (U.S. Census Bureau, 2001), generally poor performance in most domains of educational quality (Creech, 2000; Lord, 2003), and a weak system of services and supports to shore up family well-being for those who are left behind in school (Rural Poverty Research Institute).

Although there is general agreement on the seriousness of the achievement gap, there is no consensus on its causes or solutions. Some research suggests that children who attend predominantly ethnic minority or predominantly poor schools are at an academic disadvantage because of the contextual effects of social segregation (Bankston & Caldas, 1998; Entwisle & Alexander, 1992; Reardon, 2003). Other studies note the significant problems associated with persistent and severe disparities that affect the quality and resources of schools serving disadvantaged groups of students (Biddle & Berliner, 2003). A third direction of scholarship reflects a suspicion that differences in culture and family structure may be to blame for the lower school success of poor and ethnic minority youths (Murray, 1994). Assessing these differing explanations in light of social work's professional knowledge base and values is important for informing practice and policy efforts to improve the well-being and life opportunities of vulnerable children. Specifically, social work's commitment to social justice, and its conceptualization of individual functioning as intertwined with social and structural context provide an important framework for understanding and responding to achievement gaps. Toward that end, this article examines the contexts of and influences on achievement within public schools in the American South.We aim both to understand southern achievement gaps in terms of the particular educational and sociocultural context of the region and to explore the practice and policy implications of responding to achievement gaps within a social justice framework.

EDUCATION, MOBILITY, AND SOCIAL JUSTICE

Education is traditionally viewed as a leveler of opportunity. In a free and public education system, children of all backgrounds can theoretically achieve any adult status by seizing opportunities available to all and excelling based on their merit and effort (Stanton-Salazar & Dornbusch, 1995). In an unequal society with a highly residual social welfare system, however, the actual possibility of mobility through education is central to social justice efforts, creating a critical pathway to opportunity for children born into poor families and those whose families are marginalized because of racial discrimination. Problematically, a significant body of research suggests that schooling in the United States does not lead to an equalization of resources, skills, or opportunities (Braswell et al., 2001; Ferguson, 1998; Miller-Cribbs, Cronen, Davis, & Johnson, 2002). In fact, the school achievement gap between poor and nonpoor children is troublingly high (Braswell et al., 2001). Given the race-poverty overlap, it is not surprising that the poverty gap coexists with a race gap in student achievement (Jencks & Phillips, 1998).

Some scholars conceptualize achievement gaps in terms of cultural or attitudinal differences (Murray, 1994; Ogbu, 1986). For example, Murray attributed black youths' declining educational outcomes to cultural adaptations to changing incentive structures. Describing recent social policy as removing disincentives to crime, nonmarital childbearing, and welfare participation, Murray concluded that an ensuing"culture of poverty" inhibits academic success in ethnic minority communities. With a related logic, Ogbu noted an "oppositional culture" of black youths who react against mainstream expectations and disengage from school because they fear being accused of "acting white" and because they do not perceive the benefits of education. From this perspective, youths choose not to succeed in school when they are surrounded by a culture that stigmatizes achievement and when there are few material rewards to outweigh the costs of such stigma.

Other scholars agree that there are cultural dynamics involved in sustaining achievement gaps, but they focus more on the structural conditions in which local cultural norms become both necessary and logical. As Mickelson (1990) noted, "the material realities experienced by black youths challenge the rhetoric of the American Dream ... the myth that education equals opportunity for all" (p. 59). Likewise, Loury (1977) argued that structural disadvantage, in the form of inherited material and social marginalization, constrains what ethnic minority youths can achieve through equal opportunity educational programs. These constraints limit the supports, opportunities, and resources that parents from ethnic minority groups can deploy on their children's behalf. Wilson (1987) described the broad demographic and residential shifts that have isolated some ethnic minority youths in neighborhoods where few adults are employed and few parents have finished school or married before having children. This social isolation, Wilson claimed, is a significant structural barrier to academic success.

In a related vein of scholarship, some studies have focused more on the quality and value of day-to-day experiences of relationship available to youths in isolated ethnic minority or poverty milieus. Studies by Stanton-Salazar and Dornbusch (1995) and Fernandez-Kelly (1994) both found that a lack of opportunities for mentorship, relationship, support, or information from more privileged social ties forecloses many options for poor ethnic minority youths, leaving high rates of school failure and early childbearing as sadly predictable outcomes. Similarly, studies addressing peer-group effects explore the idea that a child's social ties in school influence his or her individual learning (Bankston & Caldas, 1998; Hoxby, 2000; Reardon, 2003). These studies consistently find that having a higher proportion of ethnic minority or low-income children in a school is correlated with lower levels of individual student achievement. However, such peer group effects may be expressions of culture ofpoverty dynamics if child and family risk factors accumulate at the school level due to residential segregation. From a social justice perspective, we note that even "cultural" processes may ultimately reflect structural causes to the degree that an individual's choices, beliefs, values, and behaviors are shaped by unequal access to resources and opportunities, institutional oppression, and processes of marginalization. One such structural cause is the quality of education itself.

Different and Unequal-Quality Explanations Dramatic differences in school quality are well documented, from Kozol's (1991) description of the deplorable conditions in East St. Louis to the "Corridor of Shame" depicted by Ferillo (2005) in a film documenting the inability of impoverished schools in rural South Carolina to provide even a "minimally adequate" education. But in addition to the types of obvious differences noted by Kozol and Ferillo, educational quality reflects a range of more subtle processes, experiences, and opportunities at the nexus of school and classroom environment. Within classrooms, educational quality depends on several factors: the particular qualities and attributes of the teacher, the social and physical context in which learning unfolds, and the specific activities and events structuring how children experience their time as learners.

Teachers are important as primary facilitators of the social and learning environment, and as resources, mentors, and supports for children's development. Some teacher attributes appear particularly important to predicting academic outcomes, with more experienced teachers, teachers with stronger academic and cognitive skills, and teachers with subject-specific preparation and expertise all associated with positive effects on student learning (Mayer, Mullens, Moore, & Ralph, 2000). Unfortunately, high-poverty and high-ethnic minority schools, on average, have teachers with less experience, less education, and lower levels of credentialing (Betts, Rueben, & Danenberg, 2000). As Biddle and Berliner (2003) reported, inequities in per student funding are associated with sizable differences in academic outcomes, largely because of related differences in teacher qualifications.

Along with teachers, classroom peers are important to individual student learning, as differences in race, socioeconomic status, and skill level can expose children to diverse perspectives, strengths, and norms. Although ability tracking appears to benefit high-skill students (Fertig, 2003), it limits lower-skill students' opportunities to learn from and with their more advanced classmates-with potentially damaging consequences. Pianta and colleagues (2002) documented that classrooms of predominantly low-income kindergarteners offer diminished instructional climates and teaching approaches that are less child centered. A National Center for Education Statistics (1999) study noted a greater emphasis on basic skills learning and more use of teacher-directed" routine skill" such as lecture and worksheets in classrooms within high-poverty schools. Knapp and Turnbull (1990) found that the educational approach differed depending on the percentage of students in a class who read below grade level. So in important ways, the quality and nature of the educational experience varies, depending both on a particular student's own characteristics and on the characteristics of the peers with whom that student spends classroom and school time.

With high proportions of black students and of poor white and rural students in the U.S. South, both educational equity and educational improvement are critical goals for advancing social justice and family well-being. These goals are reflected in the mandates of the No Child Left Behind Act, and the high stakes this legislation attaches to educational achievement. Achieving educational excellence among all student groups is easier mandated than it is accomplished-particularly given a strong southern conservatism that manifests itself in a minimalist social safety net (Fram, Miller-Cribbs, & Farber, 2006), general lack of support for public programs (Cooper & Knotts, 2004), and growing political pressure for school vouchers or similar mechanisms to allow privileged families to opt out of a public school system (South Carolinians for Responsible Government, n.d.). Social workers-both those working in schools and those concerned with issues of social justice and child well-being more broadly-must confront these multiple challenges to serve vulnerable children who are likely to struggle in school.The present research aims at exploring the contexts of educational achievement in the South, considering issues of race and socioeconomic status, as well as the family, classroom, and school factors that may influence learning.

METHOD

The data for this study come from the first two years of the Early Childhood Longitudinal Study Kindergarten Cohort (ECLS-K) (National Center for Education Statistics, 2001).The ECLS-K tracks the educational development of a nationally representative cohort of children, beginning with their kindergarten entry in fall of 1998. ECLS-K data are nested, with multiple children in a classroom and multiple classrooms in a school. Data were collected from various sources, including students, parents, teachers, and school administrators. The present study includes measures at the child and family, classroom, and school levels.

Our study considers a subset of the ECLS-K cohort, limited to white, black, and Hispanic students attending public school in the South who neither changed classrooms during kindergarten nor changed schools between kindergarten and first grade. These restrictions support our analytic focus and also allowed us to examine the race distinctions most salient to issues of segregation and minority status in the South. Methodologically, the restrictions with respect to classroom and school stability were chosen to reduce cross-classification in the nested data. Our resultant data included 3,501 children, in 1,208 classrooms, in 246 schools. There was an average of 15.6 children per school and 4.4 children per classroom included in the sample.

Measures

Reading Skills. Because reading is a central skill both in itself and as a tool for learning in other domains, we used it as our outcome measure throughout these analyses. Child assessments of reading skills were conducted in the fall and spring of the kindergarten year and in the spring of the first-grade year. ECLS-K reading tests were derived from the Reading Framework for the 1992 and 1994 National Assessment of Educational Progress (Rock & Pollack, 2002), and they address basic skills, vocabulary, and four types of reading comprehension skills. Assessments were scored using Item Response Theory (IRT), and we used reading IRT-scale scores at these three time points with end-of-first-grade scores as our outcome measures in the multivariable analysis.

Child and Family Variables. To better isolate school and classroom processes that may shape school achievement, we accounted for a set of child and family background and demographic factors that is also commonly thought to effect child learning. At the child level, we included gender and race (taken from the parent interview data) and the child's age at kindergarten entry (calculated from date of birth and school start date).Whether the child entered the study as a kindergarten repeater was controlled for in the multivariable models, as was full-day versus half-day kindergarten participation.

Parents act as resources for their children's learning by sharing their knowledge, investing time and energy, and acquiring material goods and opportunities that can optimize child development. Parents may be better positioned to enhance their children's learning when they have more education, more life experience, more economic resources, and the added parent-child time often possible in two-parent households.To capture some of the variability in parental contributions to children's achievement, we include indicators for mother's years of education (1 = 8th grade or less to 9 = doctorate or professional degree), family socioeconomic status (measured by income quintile), single-parent household (compared with two-parent household), and teenage childbearing (1 = yes, 0 = no). Finally, some parents are more proactive in guiding their children's learning experiences than are others. More proactive parents may take advantage of"school choice" opportunities or be highly selective about where they live to place their children in good schools, and they may also advocate for their children in other ways that we cannot measure in this dataset, but which have effects on their children's learning. To account for these parental differences, we included indicators of residential choice (Did parents chose their place of residence so their child could attend a particular school?), and nonassigned school choice (Are parents sending their child to a school they have chosen rather than to their regularly assigned school?).

Classroom Variables. To capture aspects of the classroom environment that children experienced during their first-grade year, we used data from the teacher survey administered during spring of first grade to identify three domains of classroom environment as discussed earlier.Teacher characteristics included measures of teacher's ethnicity (white or not white), teacher's years of employment in the current school, and teacher's type of certification (measured in a Likert-type scale, with higher numbers representing higher levels of credentialing). Classroom social context was indicated by the percentage of ethnic minority students in the classroom and the proportion of low reading skill students in the classroom. The physical context for learning was reflected in the averaging of teacher assessments of the adequacy of 18 classroom items (that is, books, computer equipment, classroom space), each rated from 2 = never adequate to 5 = always adequate.

Finally, we tapped aspects of educational approach, indicated by time spent in achievement groups, time spent in child-directed and teacher-directed activities, and evaluation practices based on universal versus relative standards. Our measure for achievement group time combines teachers' reports of time spent, on an average day, in reading achievement groups plus time spent in math achievement groups, and ranges from 2 = 1-15 minutes to 4 = more than 60 minutes per day. Our second classroom organization variable was the ratio of time spent in child-directed activities to time spent in teacher-directed, wholeclass activities (with range of .2 to 3).

School Variables. The South is characterized by a large black population and by high rates of rural poverty-white rural poverty in particular. To account for potential overlaps and differences associated with these conditions, this study included three school-level measures: (1) rural school status, (2) proportion of ethnic minority students, and (3) proportion of free-lunch eligible students.

DATA ANALYSES

Our first analytic step was to resolve the issue of missing data. Using SAS 9.1 and making the assumption that the data are missing at random conditional on the variables in the imputation model, we conducted multiple imputation analyses (Little & Rubin, 1987; Schafer, 1997) at the child and family, classroom, and school levels in turn.Ten data sets were imputed, and descriptive and bivariate analyses use the averages of these data sets. Hierarchical linear models made use of all imputations. Next, we examined the contexts of children's academic achievement in terms of both classroom and school characteristics. Two new dichotomous variables, high poverty school and high minority school, were created. Highpoverty school distinguishes schools with more than 50 percent free-lunch-eligible students from those with 50 percent or fewer free lunch eligible students. High-ethnic minority school distinguishes schools with more than 50 percent ethnic minority students from those with 50 percent or fewer ethnic minority students.

Finally, we estimated a series of three-level random-intercept models (Raudenbush & Bryk, 2002), explaining variability in children's readings scores in terms of school-, classroom-, and child- and familylevel characteristics. Model 1 is a baseline model, accounting for nesting at the classroom and school levels and with the only covariates being level 1 fixed effect controls for child's reading score at kindergarten entry and kindergarten exit. In model 2, a set of child- and family-level predictors were entered as level-1 fixed effects, and we calculated the resultant change in variance components, or "variance explained" (Raudenbush & Bryk, p. 74). Model 3 considers school peer group composition by adding level 3 (school level) indicators of the school's percentage of ethnic minority students and percentage of free-lunch students, as well as an indicator for rural location. It should be noted that we anticipated a smaller magnitude in school-level variability than is found in other research because by controlling for end-of-kindergarten learning at level 1, our level 3 predictors reflect the variability in first-grade classroom reading coefficients that is over and above any school-level effect on reading gains during kindergarten. In model 4, we entered level 2 (classroom level) predictors, examining the coefficients of these new variables, as well as any resultant changes in explained variability or in coefficients of previously entered variables.

One challenge to our modeling strategy was the possibility that unobserved child and family factors influence both school assignment and reading skills. This would lead to an upward biasing of estimates on school-level variables.We attempted to minimize this problem in three ways. First, to the degree that child and family factors exert a constant influence on learning, that influence should be captured in children's kindergarten scores. Controlling for kindergarten scores removed the influence associated with unobserved child and family factors from the variability in first-grade reading we are trying to explain here. Second, by including a set of child and family predictor variables, we accounted for some of the well-established factors that influence both learning and school placement. Finally, we included indicators of parents' selection of school (by residential choice and by selection of a nonassigned school) as controls in models 2 through 4.

FINDINGS

Of the children in our sample, 1,338 (38 percent) attend high-ethnic minority schools. Students attending high- or low-ethnic minority schools differ in several ways. Children with single parents disproportionately attend high-ethnic minority schools as do children whose mother became pregnant while a teenager. Children in high-ethnic minority schools also had mothers with lower levels of education, and they lived in households with lower socioeconomic status.

Classrooms in high-ethnic minority schools differ significantly from those in low-ethnic minority schools on every dimension included in this study. Some differences are likely problematic. High-ethnic minority schools have teachers with significantly fewer years at the school and lower levels of certification.Their classrooms are less adequately equipped, and they have higher proportions of students with low reading skills. Other differences are more ambiguous-perhaps problematic, perhaps just different, and perhaps reflecting appropriate and responsive efforts to meet students' needs. For example, classrooms in high-ethnic minority schools have disproportionately fewer white teachers. Furthermore, these classrooms devoted more time to achievement groups and to child-directed activities and are more likely to use universal evaluation standards.

Because of race-class overlap, as well as the high rates of white rural poverty in the South, we considered high-poverty versus low-poverty schools as well. At the child and family level, all of the patterns of difference for high-/low-ethnic minority schools also held true for high-/low-poverty schools, although the magnitude of difference varied.Thirtyfive percent of the sample attended a high-poverty school. Black students were 1.4 times as likely to attend high-poverty schools as low-poverty schools, and they were 3.8 times as likely as white students to attend high-poverty schools.

Multilevel Models

Model 1 provides a baseline analysis of the variance in each of the three levels.The intraclass correlation coefficients represent the proportion of variance in reading skills that is attributable to differences between children within classrooms (level 1), between classrooms (level 2), and between schools (level 3). In this case, we found that 79 percent of the variability in first-grade reading gains is between children, 11 percent is between classrooms, and 10 percent is between schools.

Model 2 results indicate that children who repeated kindergarten made smaller gains in reading skills, as did children from single-parent households and children of teenage mothers. Girls made greater gains in reading than did boys. Interestingly, we found no significant differences in reading gains between black and white children, or between Hispanic and white children, controlling for other variables in the equation. The addition of the set of level 1 predictors resulted in about a 3 percent increase in the explained child-level variance component, as well as a 17 percent increase in the school-level variance component.

Model 3 analyses reveal that when entered separately (analysis not shown), the percentage of ethnic minority children and percentage of free lunch children each have a significant coefficient (-.03, SE = .01 for free lunch; -.02, SE = .01 for ethnic minority).When they are both in the model along with rural location, both coefficients drop below the level of significance. Model 4 results demonstrate classroom-level variables that are significantly related to reading gains. Longer teacher tenure had a significant positive relationship to reading gains, time in child-directed activities approached significance and had a negative coefficient, and low-reading peers was significant, with a 10 percent increase in the number of low-skill readers associated with a .76 point decrease in average reading gains. Accounting for the classroom-level predictors, the coefficient on rural school increased in magnitude, becoming significant at the .05 level. The model 4 variable additions result in a 10 percent increase in explained variability at the classroom level and a 3 percent increase at the school level.

DISCUSSION

At the descriptive level, our findings suggest that school peer group composition is a significant marker of a range of differences in children's educational experiences in public schools in the South. Forty percent of sample children attended a school that had more than 50 percent ethnic minority students, and nearly as many attended a school that had more than 50 percent free-lunch-eligible students. In part, this reflects the large ethnic minority population and the high levels of child poverty in the South. However, it also is evidence of substantial concentration of disadvantaged children within a subset of public schools.

The schools into which disadvantaged children are concentrated reflect an accumulation of child and family risk factors. Along with race and income disadvantage, children in these schools have mothers with, on average, lower levels of education.The prevalence of growing up in a single-parent household and of having a teenage mother also represent potential barriers to these children's educational achievement - to the degree that these conditions may reflect less parental time and know-how for supporting children's learning.

In addition to the risk factors that children bring with them to high-poverty and high-ethnic minority schools, the classroom contexts in which they learn are in some ways different, and perhaps less desirable, than those offered to their peers in other schools. For example, less experienced and less well-credentialed teachers facilitate the learning process. Children spend more time in achievement groups, and teachers in high-poverty and high-ethnic minority schools are more likely to use universal standards for assessing students' learning. Finally, children in both high-ethnic minority and high-poverty schools are in classes where higher proportions of their classmates have below gradelevel reading skills.

Given these differences between children and classrooms in high- and low-ethnic minority and poverty schools, it is not surprising that, on average, test scores were lower in the high-ethnic minority and poverty schools.There is, as we would expect, a "gap" in achievement between these types of schools. From a social justice perspective, however, the issue is not so much whether a gap exists, but where, in the multiple layers of a child's environment, this gap is created and sustained.

In the hierarchical linear modeling (HLM) analyses we demonstrated that most of the variability in children's first-grade learning is attributable to child and family-level factors. Even when accounting for earlier learning experiences and a range of family and child characteristics, nearly 80 percent of the variability in reading is attributable to between child differences. Also important, the addition of level 1 predictors in model 2 accounts for a significant increase in explained variability at level 3. This suggests that a good deal of what appears to be between-school variability in achievement is actually attributable to clustering of child and family differences within schools.

Given social work's emphasis on social justice, it is important to ask what, fundamentally, these clustered child and family differences might mean in people's lived experiences. Traditionally, items that are measured at the individual level are thought of as behavioral choices. For example, one significant level 1 variable in model 2, "teen mother," is typically understood to represent a problematic personal choice (Murray, 1994). Some women choose to be

teenage parents, others do not. Alternatively, teenage mother can be understood as an expression of some women's lived experiences of social and economic marginality (Fernandez- Kelly, 1994). Such an interpretation makes particular sense in the South, where early childbearing is common (Lopoo, 2005) and is associated with a range of structural factors such as lack of access to adequate health care and family planning services, restrictive abortion laws, sexist cultural and religious influences, and poor educational opportunities (Fram et al., 2006). If women whose circumstances bring them to teenage motherhood also experience residential segregation such that their children attend high-poverty or high-ethnic minority schools, then broad structural disadvantage become statistically embedded within individual attributes. Advancing a more structurally oriented understanding of individual behavior is, then, one important way for social workers to bring a social justice perspective into research focusing on disadvantaged populations.

Although between-child differences dominated the models, there were nonetheless significant school- and classroom-level influences.The negative slopes on percentage ethnic minority and percentage free lunch (when entered separately) indicate that schools with higher proportions of ethnic minority or poor children have, on average, lower gains in first-grade reading. The strong overlap between race and socioeconomic composition is expressed in the nonsignificance of the coefficients when the variables are entered together. Notably, the coefficients in this study are smaller than those reported in similar research. For example, Reardon (2003) used the ECLS-K data and found much larger school peer group composition effects on first-grade reading. The difference may involve something unique about conditions and contexts of school segregation in the South. Alternatively, Reardon's analytic approach differed from this analysis, including that it did not account for classroom-level variables.

The addition of the set of classroom variables increased the explained variability at the classroom level and in the model overall. It did not, however, reduce the magnitude of the coefficients on individual-level variables or of rural school status. In fact, by reducing "noise" at the classroom level, the negative influence of rural location became stronger. This suggests that child and family, classroom, and school factors, although related, all have unique contributions to children's learning.

Among the significant classroom variables, skilllevel composition appeared to play an important role. Having higher proportions of classmates with below-grade-level reading skills lowered average student gains.The rationale behind ability tracking is that all students benefit from instruction that is tailored to their particular skill level. Along these lines, Fertig (2003) found that schools with higher levels of heterogeneity of achievement have lower levels of individual performance. Fertig went on to suggest, however, that" [ability] segregated classes might exacerbate the effect of educational and, therefore, income inequality because highly able students benefit from segregation whereas low ability students lose" (p. 15). Because a child's own learning is at the same time an individual process and a critical contextual resource for the learning of his or her peers, attention to social justice requires thoughtful balancing of the need for equity among students with the particular needs of each student.

Perhaps the most striking finding across the HLM models was the nonsignificance of race variables at all three levels. Although we found a wide range of significant differences between high- and low-ethnic minority schools at the bivariate level, the observed race difference appears to be an artifact of the overlapping of race with other dimensions of disadvantage. Once we accounted for family background factors, a child's race made no significant difference in reading gains. Similarly, once we accounted for percentage offree- and reduced-price lunch-eligible children in the school and rural status, school percentage of ethnic minority students made no significant difference. This may be attributable to the high levels of white poverty in the rural South and perhaps to the greater prevalence of teenage parenting among southern compared with non-southern white women (Lopoo, 2005). The significance of single-parent and teenage-parent variables at level 1 and of the rural variable at level 3 provided support for this hypothesis.

If, in fact, it is the greater sociostructural marginality of southern white people that accounts for the lack of "race effects," then social workers and policymakers could attend more directly to poverty and family structure as risk factors affecting student learning. Politically, however, a shift from the historical focus on race to a focus that prioritizes socioeconomic disadvantage may not be easily accomplished-not in social work and not in the South where a strong conservatism tends to eschew notions of socioeconomic status altogether.

One major limitation of the study is our limited ability to define "the South." The ECLS-K regional identifier for it is quite broad, including states with very different histories, demographics, and cultural and racial contexts. Findings of a lack of race differences and of negative influences associated with family structure and rural school location are provocative, particularly in light of mainstream assumptions of the preeminent role of black-white differences in southern culture. Future research should explore alternative definitions of the South, contrasting "Deep South" states to other southern states and perhaps treating states with high Hispanic populations separately from those with high black populations. Such analyses will be particularly important for clarifying the nature and influence of race and socioeconomic status in shaping student achievement.

IMPLICATIONS

The past decade has seen a heightened emphasis on school reform but arguably no improvement in achievement gaps (Lee, 2006). Consequently, many have begun to doubt whether the U.S. public school system is worth keeping. A movement to privatize education (South Carolinians for Responsible Government, n.d.) threatens the stability of one of our strongest universal programs. At the same time, the rolling back of social policies that support poor single-mother families and provide access to family planning and other basic services represents a threat not only to individual children's learning, but also to the overall viability of schools to meet demands for student achievement.

The movement from explaining a problem to acting to solve it is at the heart of social work's commitment to advancing social justice. The problem of educational inequality is neither new nor easily solved. This study offers one look at the complex dynamics influencing academic achievement in the South, and although our conclusions certainly warrant substantiation through additional research, we find preliminary support for several ways that social workers might promote greater equality in educational opportunities and outcomes. First, advocating for mixed-ability peer groups may empower vulnerable children toward greater school success by exposing them to high-skill peers and the expectations, opportunities, and resources afforded those students seen as having the greatest potential. Educating teachers and school administrators on the consequences of ability tracking, and building support for integration among the more privileged families whose children are overrepresented in high-skill groups are also important steps toward forging educational equity for vulnerable children. Second, our findings suggest that reducing achievement gaps will likely depend not only on educational reform, but also on a more traditional social work responsibility-strengthening the policies and programs that promote economic equality and meaningful choices about the timing and balance of human capital development, family formation, and parenting.

Finally, for those of us in the South, this study has particular implications. Although race looms large in southern understandings of daily life, we find that family structure, maternal attributes, peers' skill levels, and rural and nonrural location are more salient in explaining differences in children's achievement. Shifting attention toward these other domains of disadvantage could lead to new opportunities for progressive change by pointing to the common needs of children of all races. Such a shift, we hope, can lay a foundation for galvanizing a broader, shared commitment not only to effective schools, but also to more adequate social programs and economic policies that would promote family life and enhance community well-being." (Maryah Stella Fram, Julie E. Miller-Cribbs, and Lee Van Horn, Social Work, October 2007)